International Journal of Healthcare Management

ISSN: 2047-9700 (Print) 2047-9719 (Online) Journal homepage: http://www.tandfonline.com/loi/yjhm20

Healthcare financing in South-East Asia: Does

fiscal capacity matter?

Deepak Kumar Behera & Umakant Dash

To cite this article: Deepak Kumar Behera & Umakant Dash (2018): Healthcare financing in
South-East Asia: Does fiscal capacity matter?, International Journal of Healthcare Management,
DOI: 10.1080/20479700.2018.1548159

To link to this article: https://doi.org/10.1080/20479700.2018.1548159

Published online: 25 Nov 2018.

Submit your article to this journal

Article views: 14

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=yjhm20
INTERNATIONAL JOURNAL OF HEALTHCARE MANAGEMENT
https://doi.org/10.1080/20479700.2018.1548159

Healthcare financing in South-East Asia: Does fiscal capacity matter?

Deepak Kumar Behera and Umakant Dash
Department of Humanities and Social Sciences, Indian Institute of Technology Madras, Chennai, India

ABSTRACT ARTICLE HISTORY

Financing healthcare and achieve sustainable development goals by 2030 is the prime focus of Received 20 June 2018
many low-income and middle income countries. Improved public healthcare and reduce the Accepted 9 November 2018
burden of out-of-pocket health expenditure is the overarching public health policy objective
KEYWORDS
across the Asia-Pacific region. This study examines the impact of fiscal capacity on health Healthcare financing; South-
expenditure by controlling socio-economic factors in South-East Asia for the period of 1995– East Asia; fiscal capacity;
2013. We have employed a panel fixed effects regression model in order to capture country- Health expenditure; fixed
level unobserved heterogeneity and macroeconomic policy changes in the health sector. effects model
Overall result shows that annual change in per capita government health expenditure
(pooled financing) is influenced positively by the per capita income, fiscal capacity, ageing, JEL
and the prevalence of tuberculosis and rate of urbanization. The income elasticity of I130; H510; H3; C330; H510
government health expenditure is less than one, which implies that the health expenditure is
treated as a necessity for the public health providers. Overall empirical analysis concludes
that pooled financing mechanism positively influenced by the size of fiscal capacity thereby
the share of out-of-pocket health expenditure to total health expenditure declines over the
period. The result suggests that the faster movement towards health financing transition
would possible through the generation of fiscal capacity by improving public finance policies.

1. Introduction
and tested whether the health financing transition is
Improved public healthcare and reduce the burden of observed on average in SEAR during this relatively
out-of-pocket health expenditure is the overarching shorter-time period.
public health policy objective across the Asia-Pacific This study has selected the SEAR due to the follow-
region [1]. The World Health Organization (WHO) ing four reasons. First, SEAR capture not only to a
framework on the health system1 argues that financing quarter of the world’s population but also to a daunting
healthcare plays an influential role for the attainment range of both communicable and non-communicable
of Universal Health Coverage (UHC2) and reduce the diseases. The health status of the world is significantly
financial burden of households by protecting them influenced by the health status of the SEAR [1]. Second,
against the costs of illness [2]. In this study, we have the distinctive characteristics of this region are that
adopted health financing framework as suggested by there is a huge diversity in terms of geography, linguis-
Fan and Savedoff [3]. They have argued that most tic, and political structure but these countries generally
low-income and middle-income countries show two face common health challenges. The countries like
basic health financing trends over time. First, per capita Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan,
total health expenditure increases due to annual incre- and Sri Lank face the triple burden of persisting infec-
ment in the share of public financing towards health- tious diseases, incomplete demographic transition,
care. Second, total out-of-pocket health expenditure HIV and AIDS, massive unplanned urbanization.
as a share of total health expenditure declines due to Further, the health system of SEAR is very weak such
the incremental change in the resource mobilization as limited national health insurance schemes and a lar-
capacity of economies. Therefore, they have described ger role of the private providers of health payments [2].
the trends and patterns of health expenditure move- Third, SEAR has gained substantial improvement in
ment over the time period as the health financing tran- health-related millennium development goals from
sition. By following this argument, we have examined 2000 to 2013 such as reduction of under-five mortality
the trends and patterns of health financing transition rate and maternal mortality ratio, reduction in the
movement in the South-East Asia Region (SEAR) incident of malaria and HIV but it has been limited
such as Bangladesh, Bhutan, India, Indonesia, Mald- progress in the percentage of antenatal coverage and
ives, Myanmar, Nepal, Sri Lanka, and Thailand for access to sanitation facilities. The countries such as
the period 1995–2013. Basically, this study has empiri- Bhutan, Indonesia, Maldives, Sri Lanka and Thailand
cally examined the determinants of health expenditure have made good progress in placing primary healthcare

CONTACT Deepak Kumar Behera deepakkumar.behera59@gmail.com

© 2018 Informa UK Limited, trading as Taylor & Francis Group
2 D. K. BEHERA AND U. DASH

oriented health systems in order to strengthen the goal expenditure trends in SEAR, we have observed three
of UHC while other countries of this region are now important patterns of health financing transition by
preparing for success [5]. Fourth, as the best of author’s comparing the annual change in pooled (government)
knowledge, the examination of the determinants of per capita health expenditure with the annual change
health financing transition (four composition- per in out-of-pocket health expenditure (Figure 2). Pattern
capita total health expenditure, per capita government 1: When pooled health expenditure rise and out-of-
health expenditure, per capita out-of-pocket health pocket expenditure declines or stays the same,
expenditure and out-of-pocket as percent of total countries move rapidly through the health financing
health expenditure) and their movement over the transition (countries such as Thailand, Maldives, and
period has unnoticed by the existing literature. Bhutan). Pattern 2: When pooled health expenditure
The study has applied panel fixed effects regression rises faster than the pace at which out-of-pocket expen-
model in order to examine the determinants of health diture rises, countries progress through the health
financing transition by controlling economic, political, financing transition more slowly (countries such as
and demographic factors. Overall result shows that India, Nepal, Sri Lanka, and Indonesia). Pattern 3:
annual change in per capita government health expen- When pooled health expenditure grows more slowly
diture (pooled financing) is influenced positively by an than the out-of-pocket expenditure, countries regress
economic factor (per capita income), political factor (countries such as Bangladesh and Myanmar).
(fiscal capacity), demographic factor (ageing), the In Figure 3, we have examined the sources of health
prevalence of disease (Tb) and urbanization in the financing in SEAR for the period of 1995–2013. WHO
SEAR for the period of 1995–2013. [1] has proposed two important health financing tar-
This paper is organized as follows. Section 2 con- gets in order to attain UHC in the Asia-Pacific region.
tains a brief overview of health financing trends and First, out-of-pocket health expenditure should not be
patterns in the SEAR. Section 3 reviews the existing lit- exceed the 30 percent-40 percent of total health expen-
erature. Section 4 discusses the data and methods. Sec- diture; second, over 70 percent of the population
tion 5 discusses the empirical results. Section 6 contains should be covered by prepayment and risk pooling
the conclusion. schemes. We have found three important insights
from Figure 3. First, countries such as Bangladesh,
India, Indonesia, Myanmar, and Nepal have shown a
2. Brief overview of health financing trends
higher out-of-pocket expenditure (above the threshold
in South-East Asia
limit of 30 percent-40 percent) associated with a lower
Healthcare financing plays a critical role in progressing government health expenditure. The result implies that
towards the goal of UHC in SEAR but the region these countries provide less financial protection to their
suffers huge financial constraint resulting in inadequate people and limited coverage of prepayment schemes.
public investment in health. In other words, govern- Second, countries such as Bhutan, Maldives, and Thai-
ment revenue which mediates the allocation of health land have shown a lower out-of-pocket expenditure
expenditure is limited in this region. As a result of (below the threshold limit of 30 percent-40 percent)
which there is high and even improvising out-of- associated with a higher government health expendi-
pocket expenditure by households [5]. ture. It implies that these countries have been provided
Fan and Savedoff [3] argues that in most low- financial coverage under various prepayment mechan-
income and middle-income countries, out-of-pocket isms. Third, countries like Indonesia, Maladies, and
expenditure increases in absolute terms but its share Thailand have raised their total health expenditure
of total health expenditure declines because pooled through private insurance while other SEAR countries
expenditure grow faster. In this study, we have just started their dependency through private insur-
observed a similar pattern of transition in the context ance mechanisms.
of SEAR. Figure 1 presents trends in the health finan- In this paper, we have also examined why pooled
cing transition – the increase in total health expendi- health financing of countries such as Thailand, Bhutan,
ture on the left y-axis and the decline of out-of-pocket and the Maldives shows a faster movement towards
expenditure (OOPE) as a percent of total health health financing transition than the other SEAR
expenditure (OOPE%THE) on the right y-axis in countries. Therefore, we have shown the fiscal capacity
SEAR. The result implies that total health expenditure trends of SEAR. Existing literature includes Heller [6],
has been increased and the share of OOPE%THE Durairaj and Evans [7], and Mclntyre and Kutzin [8]
declined in between 1995 and 2013. argue that fiscal capacity (i.e. the overall level of
Further, Fan and Savedoff [3] has argued that it is government expenditure) shows the current size of
not necessary for all countries move towards health the public sector in the economy. They suggested cer-
financing transition in the same speed over the time tain “rule of thumb” for assessing the fiscal capacity
and would be influenced by the public policy of the of an economy such as low fiscal capacity (<20%), med-
government. In order to examine the health ium (20%-35%), and high (>35%).
 INTERNATIONAL JOURNAL OF HEALTHCARE MANAGEMENT 3

Figure 1. Total health expenditure and out-of-pocket as a share of total health expenditure, 1995–2013.

Figure 4 shows that countries like Bhutan and Schiber and Maeda [15]; Musgrove et al. [16] have
Maldives have been a high level of fiscal capacity; categorized countries in different income panels.
countries like India, Sri Lanka, and Thailand have They find that the share of pooled health financing
been a medium level; and countries such as Bangla- (government health expenditure) to income moving
desh, Indonesia, Myanmar, and Nepal have been a very slowly towards health financing in the lower
low level of fiscal capacity. We have found two middle-income countries, while the share of pooled
insights. First, despite the medium level of fiscal health financing increases rapidly and share out-of-
capacity, the fiscal capacity for health (ratio of govern- pocket expenditure decreases in high-income
ment health expenditure to general government countries.
expenditure) is larger in Thailand than the Bhutan Earlier literature has emphasised that per capita
and Maladies. Second, fiscal capacity is one of the income is the ultimate source of health sector financing
potential sources of health sector financing in SEAR. and they have given less importance to the role of pub-
Overall trends analysis of health financing transition lic finance policies on health financing. In the emer-
concludes that pooled financing mechanism positively gence of United Nations Sustainable Development
influenced by the size of fiscal capacity thereby the Goals and international pressure for achieving health
share of out-of-pocket health expenditure to total financing coverage in low-income and middle-income
health expenditure declines over the period. countries, the role of fiscal resources became crucial
to ensure better access to essential healthcare services
and financial protection [4,7]. Fan and Savedoff [3],
3. Review of literature
Hitiris [17], Hartwig and Sturm [18], Ke et al. [19],
There has a bulk of literature examines the determi- and Behera and Dash [20] have used government
nants of total health expenditure and find that rising expenditure as percent of GDP (fiscal capacity) as
income is the major source of health expenditure one of the explanatory variables to explain the determi-
growth [9–13]. They argue that if the income elasticity nants of government expenditure on health . They have
of health expenditure is more than 1 percent, then we found that the fiscal capacity shows a positive and stat-
assume that the incremental change in healthcare istical significant relationship with the per capita gov-
expenditure is more than the incremental change in ernment health expenditure in low-income and
per capita income and healthcare treated as a luxury. middle-income countries. Fiscal capacity measure the
On the other hand, if the income elasticity of health resource mobilization capacity of the government
expenditure is less than 1 percent, then we assume towards healthcare financing that leads to reduce the
that the incremental change in healthcare expenditure share of out-of-pocket health expenditure to total
is less than the incremental change in per capita health expenditure [3,4,19].
income and healthcare treated as a necessity. They The increasing share of older people to the total
argue that income elasticity is less than 1 percent that population called ageing and the effect of ageing on
implies healthcare expenditure increases more slowly the growth of government health expenditure as well
than income and suggests that the government needs as the growth of out-of-pocket health expenditure is
to intervene in health system financing. There is a pau- the subject of scholarly debate among the health
city of literature who examines the determinants of the researchers because of its mix impact on the growth
total health expenditure and its composition using of total health expenditure. Lai [13], Crivelli et al.
panel data econometric methods. Farag et al. [14]; [21]; Cantarero and Lago-Penas [22]; Khan and
4 D. K. BEHERA AND U. DASH

Figure 2. Changes in pooled and out-of-pocket health expenditure, 1995–2013.

Mahumud [23]; find that the large share of older move towards universal health coverage. Crivelli et al.
people (age on and above 65) tends to increase health- [21], Cantarero and Lago-Penas [22], Farag et al.
care costs. They argue that an older population tends to [14], and Lai [13] have included ‘time’ trend as proxy
cause higher health expenditure because of the increase for changes in medical technology and captures the
incidence of illness as well as proximity to the time of cost difference over time due to changes in medical
the death of the elderly. While Fan and Savedoff [3], technology or other factor that may affect the growth
Barros [23], Sen [24], find that ageing show an of health expenditure at the national level. The techno-
insignificant impact on total health expenditure. logical change increases in the medical care cost over
Cutler and McClellan [25], Fan and Savedoff [2], time, it does not necessarily mean that technological
De Meijer et al. [26], Sorenson [27], find medical tech- change is bad because often bring health improvement
nology, rising wages of health personnel, and prices of and it seems to bad only if the cost increases more than
the medicine are increased healthcare cost and influ- the benefits.
ence the financial allocation of the government. Their Nistico et al. [29] argue that reduction of epidemics
argument is that changes in medical practices and tech- of infectious diseases improves the quality of life of the
nological innovations in terms of replacing traditional people and increase the life expectancy. They argued
drugs, diagnostics and procedures have led to an there is a positive correlation between socio-economic
increase the cost of medicine. Mohapatra and Murarka vulnerability and mortality due to the prevalence of
[28] argues that technology up-gradation reduced the diseases. They suggested that healthcare system should
cost of healthcare and improved the quality of health- consider socio-economic factors before financing uni-
care services. They suggested that health financing versal healthcare in an economy. Blanco-Moreno
strategy should prioritize the resources allocation by et al. [30] argue that death-related cost diminishes as
judging the economic scale of the decision making health status improves and intensity of healthcare use
units such as hospital and health center in order to is the main driver of health expenditure in Spain.

1995-2013
90
as % of Total health expenditure

80
70
60
50
40
30
20
10
0

Government health expenditure Out-of-pocket expenditure Private insurance

Figure 3. Source of health financing in SEAR, 1995–2013.

 INTERNATIONAL JOURNAL OF HEALTHCARE MANAGEMENT 5

40

35

30

25

Percent 20

15

10

0
h

ka
ar
s

nd
n

al
ve
es

si
ta

di

nm

ep

an

la
ne
ad

hu

In

di

ai
iL
ya
do

al
gl

Th
M

M
an

Sr
In
B

GGE as % of GDP (Fis cal capacity)

GHE as % of GGE (Fis cal capacity for health)
Ratio (GHE/GGE)

Figure 4. Trends of Fiscal capacity towards health financing in SEAR, 1995–2013.

They suggest that keeping expenditure under control Asia Region (SEAR) countries such as Bangladesh,
country should be focused on factors such as popu- Bhutan, India, Indonesia, Maldives, Myanmar, Nepal,
lation’s health status, economic growth and develop- Sri Lanka, and Thailand for the period of 1995–2013.
ment, new technologies and medical progress, and The countries such as the Democratic People’s Repub-
the organization and management of the healthcare lic of Korea and Timor-Leste are excluded due to non-
system. Dickov and Dickov [31] argue that health sys- availability of data of entire study period. We have
tem management and organizational strategy creates included four types of health financing indictors3
the investment environment in the healthcare sector such as per capita total health expenditure (PCTHE),
both at the global and the local level. They argue that per capita government health expenditure (PCGHE),
health system required auditing of financial statements per capita out of pocket health expenditure (PCOOPE),
by healthcare institutions at the local level and the and out-of-pocket expenditure (OOPE) as a percent of
operational audit of economic policy in the healthcare total health expenditure (OOPE%THE). Mclntyre and
institutions at the national level. Reddock [32] argues Kutzin [8] argued that the assessment of these health
that supply-side and demand-side perspective should finance indicators is critical for moving towards
be included in the assessment of financial coverage of UHC. They argued that the PCGHE provides insights
healthcare cost, thereby it could reduce the share of into the level of government expenditure on health
out-of-pocket health expenditure. They suggest that and it is the strong predictor of the extent to which
the supply-side and demand-sides of health services the health system depends on OOPE. On the contrary,
are complementary each other that should be con- OOPE%THE is of critical importance in assessing the
sidered before designing the healthcare financing fra- extent of financial protection within a country.
mework for UHC. Hence, reducing the share of OOPE to total health
expenditure and increasing the pooled government
health expenditure (PCGHE) should be a policy pri-
4. Data and methods
ority of many countries for achieving UHC.
This study has examined the elasticity of health expen- Table 1 shows the result of descriptive statistics and
diture with respect to fiscal capacity by controlling pair-wise correlations. We have obtained the data of
other pertinent factors such as per capita Gross Dom- variables for the period of 1995–2013 from the Global
estic Product (PCGDP), ageing, the prevalence of Health Expenditure Database of WHO [33] and World
Tuberculosis (Tb) and rate of urbanization in SEAR. Development Indicator of World Bank [34]. The mean
PCGDP shows the level of income of a country, fiscal value of PCTHE is 194.39 US$, PCGHE is 107.82 US$,
capacity shows the government’s ability and willing- PCOOPE is 68.69 US$ and OOPE %THE is 47.79 per-
ness to spend on healthcare; ageing shows the demo- cent respectively. So, on an average of around 48 per-
graphic structure of the country; the prevalence of cent of health expenditure is sourced from
Tuberculosis shows the morbidity situation, and rate household’s finance which seems to be a financial bur-
of urbanization shows the demand for healthcare ser- den due to huge medical expenses. The higher standard
vices. In this study, we have selected nine South-East deviation (Std. Dev.) result shows that there is huge
6 D. K. BEHERA AND U. DASH

Table 1. Result of descriptive Statistics and Pair-wise correlations.

Descriptive Statistics Pair-wise correlations
Variable Mean Std. Dev. Min Max PCTHE PCGHE PCOOPE OOPE % THE
PCTHE 194.39 180.28 8.76 1124.34 1
PCGHE 107.82 123.98 0.98 641.48 0.960*** 1
PCOOPE 68.69 57.09 7.45 426.56 0.814*** 0.631*** 1
OOPE % THE 47.79 20.7153 11.22 90.64 −0.652*** −0.742*** −0.256*** 1
PCGDP 4393.99 3004.2 444.24 12418.77 0.814*** 0.792*** 0.596*** −0.707***
Fiscal Capacity 23.42 9.23 3.52 49.72 0.571*** 0.596*** 0.397*** −0.618***
Ageing 5.07 1.34 3.14 9.71 0.321*** 0.333*** 0.198*** −0.251***
Tb 198.58 112.82 38 563 −0.527*** −0.411*** −0.545*** 0.398***
Urban 29.10 9.47 10.9 52.25 0.429*** 0.427*** 0.291*** −0.285***
Notes: Balanced panel of nine countries such as Bangladesh, Bhutan, India, Indonesia, Maldives, Myanmar, Nepal, Sri Lanka, and Thailand over the period
1995–2013. PCTHE: Per capita total health expenditure, PCGHE: Per capita general government health expenditure, PCOOPE: Per capita out-of-pocket
health expenditure, OOPE % THE: Out-of-pocket health expenditure as percent of total health expenditure, PCGDP: Per capita gross domestic product,
Fiscal Capacity: General government expenditure as percent of GDP, Ageing: Population ages 65 and above as percent of total population, Tb: Prevalence
of tuberculosis per 100,000 people, and Urban: Urban population as percent of total population. GDP and health expenditure are in million constant 2005,
US$. PCTHE, PCGHE, PCOOPE, PCGDP are in US$. *** denotes level of significant at 1%.
Source: Author’s estimation from the Global Health Expenditure Data Base of WHO [33] and World Development Indicators of World Bank [34].

variability in economic indicators such as fiscal estimators because cross-sectional unit has its own
capacity and per capita GDP in SEAR. Further, there constant term and it varies within the cross-sectional
is a huge disparity in health financing indicators such unit. While the individual country-specific intercept
as PCGHE and PCTHE by comparing the minimum vi is not constant over time rather random, called as
and maximum values in SEAR. The minimum value random effects (RE) model. The RE model assumes
of OOPE%THE is 11.22 percent and the maximum that vi are uncorrelated with regressors. If the regres-
value is 90.64 percent which shows the ranges of the sors are correlated with vi and they are correlated
data over the period. The pair-wise correlations results with the composite error term (vi + 1it )and RE estima-
illustrate three preliminary finding before doing any tor will be inconsistent [35].
impact analysis through regression methods. First, The appropriateness of using FE and RE model in
there is positive relationships between PCTHE, the empirical estimation will be verified through the
PCGHE, PCOOPE and PCGDP; second, there is a Hausman test. The Hausman test shows whether the
negative relationship between OOPE%THE and fiscal regressors are correlated with the vi (in case of FE) or
capacity; Third, there is a negative relationship between uncorrelated with the vi (in case of RE). The null
PCGHE and OOPE%THE. Hence, these results pro- hypothesis of the Hausman test says there is no sys-
voke us to examine the extent and nature of relation- tematic difference in the coefficient of FE and RE esti-
ships between variables using advanced econometric mation. The rejection of the null hypothesis implies
methods. that there is a systematic difference in the coefficient,
This study has applied panel fixed effects (FE) model so we need to apply the fixed effects model rather ran-
to examine the impact of fiscal capacity on health dom effects model and vice versa [36]. In this study, we
expenditure by endogenizing per capita income, age- have estimated the fixed effects model after the rejec-
ing, Tb and urbanization in SEAR for the period of tion of the null hypothesis through the Hausman test
1995–2013. We formulated the following panel and aim of the FE model is to eliminate vi to give
regression methods. efficient estimators.
One-way fixed effects Eq. (1) shows, the one-way FE model without time
dummy and Eq. (2) shows the two-way FE model
yit = ai + bx′it + vi + 1it (1) with time dummy along with other explanatory vari-
ables. The one-way FE model which controls for
Two-way fixed effects
time-invariant country-specific unobservable effects.
yit = ai + bx′it + gt + vi + 1it (2) The two-way FE model is very relevant because
there are time trends and year specific shocks such
For i = 1, . . . , N and t = 2, . . . , T, Where as war, epidemics, droughts which could be biased
yit is the dependent variable; xit is the ((k − 1) × 1) vec- our empirical estimation [14]. In Eq. (2) the coeffi-
tor of strictly exogenous explanatory variables; vi is an cient g will measure an annual rate of change over
unobserved individual effect/homoscedastic country- time ′ t ′ in the dependent variables. This time trend
specific effects; and 1it is an unobserved white noise measures any consistent annual changes of some
disturbance/the stochastic disturbance term. The FE unobservable factors such as changing technology
model assumes that the country effects vi is constant and medical practices, cost pressures and public pol-
over time and space while the slope estimates b are icies, may contribute to the growth of health
constrained across units. It is also called within expenditure.
 INTERNATIONAL JOURNAL OF HEALTHCARE MANAGEMENT 7

5. Empirical results and discussion both public providers through government financing
healthcare and private providers in terms of house-
Table 2 represents the results of panel FE model which
hold’s medical expenditure are simultaneously contrib-
estimates the elasticity of health expenditure (i.e.
uted towards rising overall health expenditure in
PCTHE, PCGHE, PCOOPE, OOPE%THE) with
SEAR.
respect to per capita GDP (per capita income) by con-
trolling fiscal capacity, ageing, prevalence of Tubercu-
losis and rate of urbanization. Table 2, we have 5.2. Health expenditure and fiscal capacity
shown the result of both panel one-way FE (without
Table 2 examines the impact of fiscal capacity on health
time trend) and panel two-way FE (with time trend)
expenditure (i.e. PCTHE, PCGHE, PCOOPE, and
models for robustness of our results.
OOPE%THE). The result shows that fiscal capacity
shows a positive and statistically significant relation-
ship with per government health expenditure
5.1. Health expenditure and income
(PCGHE). It implies that at 1 percent incremental
Table 2 shows that per capita GDP (PCGDP) shows change in fiscal resources leads to 0.015 percent change
positive and statistically significant relationships with in the public sources of financing towards healthcare.
PCTHE, PCGHE, and PCOOPE in both one-way FE On the other hand, the fiscal capacity shows a negative
and two-way FE regression model. The result implies and statistically significant relationship with OOPE%
that at 1 percent incremental change in PCGDP, total THE. It implies that at 1 percent incremental change
health expenditure (PCTHE) increases at less than 1 in the resource mobilization leads to 0.29 percent
percent and the coefficient of PCTHE is 0.76 percent reduction in OOPE%THE. This result exhibits two
(column 2). The regression result of per capita govern- major insights. First, the mobilization of resources
ment health expenditure (PCGHE) shows that at 1 per- through increasing overall government expenditure
cent incremental change in PCGDP leads to 0.65 would be capable of reduce the out-of-pocket health
percent change in PCGHE and the elasticity is less expenditure. Second, fiscal capacity has a marginal
than 1 percent (column 4). The regression result of impact on pooled sources of financing and it needs to
per capita out-of-pocket expenditure (PCOOPE) be improved further. Our results are similar to those
shows that at 1 percent incremental change in of Hartwig and Sturm [18], Ke et al. [19], Lu et al.
PCGDP leads to 1.09 percent change in PCOOPE [38], Palangkaraya and Yong [39], Hitiris [17], Gupta
and the elasticity is greater than 1 percent (column and Mondal [4], Behera and Dash [20], who find a
5). On the other hand, the elasticity of PCOOPE with positive and significant relationships between govern-
respect to PCGDP is less than 1 percent after control- ment health expenditure and fiscal capacity. They
ling both country and year effects (column 6). The argued that the government expenditure as a ratio of
result implies that at 1 percent incremental change in GDP shows the size of government and the ability of
PCGDP leads to 0.59 percent change in PCOOPE (col- government to finance healthcare. We find that the
umn 6). Overall results show that per capita income fiscal capacity shows an insignificant impact on per
growth is the most important factor in order to finan- capita out-of-pocket health spending and the result is
cing healthcare. obvious because fiscal capacity is the part of public pro-
The economic interpretation of these findings is that viders of healthcare services. So, our result proves an
the elasticity of health expenditure with respect to inverse relationship between the annual change in
income is equal to or greater than 1 percent indicating PCGHE and annual change in the share of OOPE%
that healthcare is a luxury rather than a necessity. THE that follows the health financing transition argu-
When the elasticity is less than 1 percent, healthcare ment of Fan and Savedoff [3].
is closer to being a necessity and hence needs more
government intervention. Our results find that health-
5.3. Health expenditure and ageing
care expenditure is treated as a necessity in SEAR
because the elasticity of both total health expenditure In Table 2, we have regressed health financing indi-
and composition of health expenditure are less than 1 cators (i.e. PCTHE, PCGHE, PCOOPE, OOPE%
percent. This result is similar to other studies such as THE) with the population ages 65 and above as a per-
Sen [25], Cantarero and Lago Penas [22], Baltagi and cent of total population (ageing). We find two interest-
Moscone [12], Ke et al. [19], Farag et al. [14], Fan ing results. First, there is a positive and statistically
and Savedoff [3] and Khan and Mahumud [23], and significant relationship between PCGHE and ageing
Behera and Dash [20,37]. They have also used the and the result implies that at 1 percent increment in
panel fixed effects model and find the elasticity of the share of the aged population to the total population,
health expenditure with respect to income is less than leads to 0.06 percent change in government expendi-
1 percent and the coefficient varies in between 0.40- ture on health. Second, there is a negative and statistical
0.89 percent. Therefore, our study concludes that significant relationship between out-of-pocket health
8 D. K. BEHERA AND U. DASH

Table 2. Result of panel fixed effects model.

ln(PCTHE) ln(PCGHE) ln(PCOOPE) OOPE % THE
One-way Two-way One-way Two-way One-way Two-way One-way Two-way
Variables 1 2 3 4 5 6 7 8
ln (PCGDP) 0.846*** 0.763*** 0.738*** 0.651*** 1.093*** 0.595** 2.464 0.708
(0.058) (0.123) (0.082) (0.169) (0.118) (0.239) (2.707) (5.549)
Fiscal Capacity 0.002 0.002 0.014*** 0.015*** −0.003 0.001 −0.319** −0.294**
(0.002) (0.003) (0.003) (0.004) (0.005) (0.005) (0.124) (0.135)
Ageing −0.049* −0.064* 0.068* 0.057 −0.272*** −0.357*** −6.309*** −6.624***
(0.026) (0.032) (0.037) (0.045) (0.053) (0.063) (1.217) (1.482)
ln (Tb) 0.370*** 0.335*** 0.385*** 0.359*** 0.124 −0.093 −9.716*** −10.79**
(0.080) (0.093) (0.112) (0.129) (0.160) (0.182) (3.677) (4.220)
Urbanisation 0.035*** 0.032*** 0.034*** 0.031*** 0.025** 0.006 −0.018 −0.098
(0.005) (0.006) (0.007) (0.009) (0.010) (0.012) (0.232) (0.297)
Constant −4.734*** −3.840*** −5.734*** −4.854** −4.774*** 0.860 117.5*** 139.3**
(0.660) (1.383) (0.922) (1.901) (1.325) (2.686) (30.35) (62.27)
Country fixed effects Yes Yes Yes Yes Yes Yes Yes Yes
Year fixed effects No Yes No Yes No Yes No Yes
Hausman fixed [41.22]*** [12.26]** [14.44]** [45.35]***
R2 (within) 0.864 0.870 0.817 0.830 0.550 0.595 0.292 0.347
F-test 77.49*** 66.47*** 136.86*** 131.95*** 21.22*** 16.86*** 69.64*** 65.05***
Number of country 9 9 9 9 9 9 9 9
Observation 171 171 171 171 171 171 171 171
Notes: Standard errors in parentheses; Chi (square value) in square brackets; *** indicates significant at 1 percent level; ** indicates significant at 5 percent
level; * indicates significant at 10 percent level; One-way: Country fixed effects model, Two-way: Country and year fixed effects model; ln: Natural logarithm.
Source: Author’s estimation.

expenditure and ageing population and the result to 6.62 percent reduction in out-of-pocket expenditure
implies that a 1 percent increment in the share of the as ratio a of total health expenditure. Further, there is a
aged population to the total population, leads to positive relationship between per capita total health
0.035 percent reduction in household expenditure on expenditure and prevalence of Tb, which implies that
healthcare. The economic interpretation of a negative rising government health expenditure in order to pro-
association between out-of-pocket health expenditure vide financial incentives in the disease programmes
and ageing implies that above 65 age people do have leads to raise overall health expenditure in the SEAR.
the income to spend on healthcare and they depend
on earning members of the family. Further, the result
shows a reduction of total health expenditure that is 5.5. Health expenditure and urbanization
because the impact of aging is more on the reduction In Table 2, we have regressed health financing indi-
of out-of-pocket expenditure and less on the rise of cators (i.e. PCTHE, PCGHE, PCOOPE, OOPE%
government health expenditure. Our result is similar THE) with the proportion of urban population to the
to Sen (2005) that finds that ageing shows a negative total population (Urbanization). We find that urbaniz-
relationship with per capita out-of-pocket health ation shows a positive and statistically significant
expenditure. On the contrary, our result is similar to relationship with per capita government health expen-
those such as Crivelli et al. [21], Cantarero and Lago- diture and its contribution more towards the rising in
Penas [22], Khan and Mahumud [23] and Di Matteo total healthcare expenditure in the SEAR. It implies
[11] who find that large share of old people tends to that growing urbanization leads to an increase in
increase healthcare costs thereby increases government healthcare demand such as a number of hospitals, a
expenditure on healthcare. number of medical staff etc. The result shows at 1 per-
cent increase in the urban population leads to 0.03 per-
cent change in per capita government health
5.4. Health expenditure and disease pattern
expenditure. On the other hand, urbanization shows
In Table 2, we have regressed health financing indi- a negative relationship with OOPE% which implies
cators (i.e. PCTHE, PCGHE, PCOOPE, OOPE% that rising urbanization promotes accessibility of pub-
THE) with the prevalence of tuberculosis per 100,000 lic healthcare services and reduces household’s out-of-
people (Tb). We find two interesting results. First, pocket health expenditure. Our results similar to Gerd-
there is a positive and statistically significant relation- tham et al. [10] who finds that urbanization shows a
ship between PCGHE and prevalence of Tb and the positive and significant relationship to government
result implies that at 1 percent increment in the share health expenditure.
of Tb patient, leads to 0.35 percent change in govern- The overall result of Table 2 shows that annual
ment expenditure on health. Second, there is a negative change in per capita government health expenditure
and statistically significant relationship between OOPE (pooled financing) is influenced positively by an econ-
%THE and prevalence of Tb and the result implies that omic factor (per capita income), political factor (fiscal
a 1 percent increment in the share of Tb patient, leads capacity), demographic factor (ageing), prevalence of
 INTERNATIONAL JOURNAL OF HEALTHCARE MANAGEMENT 9

disease (Tb) and urbanization in the SEAR over the sector growing demands as well as reduced the out-
period 1995-2013. While, the annual change in the of-pocket health expenditure. Although this study
share of OOPE%THE is influenced negatively by makes a preliminary attempt to understanding the
fiscal capacity, ageing, and Tb. effect of fiscal capacity on various components of
health expenditure for SEAR by controlling socio-
economic and demographic factors. This study could
6. Conclusion and policy implications
also serve as a basis for showing useful directions for
The objective of this study was to observe the pattern of carrying out similar studies for other regions of Asia-
health financing transition in SEAR over the period Pacific and other regions of WHO. Further to date,
1995–2013 and examine the impact of fiscal capacity empirical work on determinants of health financing
on the growth of health expenditure by endogenizing transition has been sparse. Therefore, this work will
per capita GDP, ageing, the prevalence of tuberculosis, be helpful for designing effective public finance policies
and urbanization. We observed that Thailand is mov- for achieving UHC in low-income and middle-income
ing rapidly through the health financing transition in countries.
which per capita government health expenditure
increases while per capita out-of-pocket health expen-
Notes
diture decline rapidly as showing negative. In a low-
income country like Nepal is moving towards the 1. Health system includes all the activities whose primary
health financing transition in the overall period. The purpose is to promote, restore or maintain health. In
precisely health systems are not just concerned with
lower middle-income countries such as Bhutan,
improving people’s health but with protecting them
India, Indonesia, Myanmar, and Sri Lanka are moving against the financial costs of illness. The challenge
towards the health financing transition at the slower facing governments in low-income countries is to
pace. reduce the regressive burden of out-of-pocket pay-
The empirical results show that annual change in ment for health by expanding prepayment schemes,
per capita government health expenditure (pooled which spread financial risk and reduce the spectra of
catastrophic health care expenditures [2].
financing) is influenced positively by an economic fac- 2. UHC provides assurance of health services to all needy
tor (per capita income), political factor (fiscal capacity), people under three objectives such as equity in access,
demographic factor (aging), a prevalence of disease quality of health services and ensuring financial risk
(Tb) and urbanization in the SEAR. While, the annual protection [2].
change in the share of out-of-pocket health expendi- 3. Total health expenditure (THE) included both public
and private expenditure on health both domestic
ture to total health expenditure is influenced negatively
and external agents. Government health expenditure
by fiscal capacity, aging, and Tb. By comparing one- from domestic sources (GHE) included government
way and two-way fixed effects regression model, the expenditure on health from general government rev-
result shows the growth of health expenditure and enue and payroll taxes. It excluded External funds
movement towards health financing transition channeled through governments. Out-of-pocket
influenced by the country-specific individual factor as expenditure (OOPE) included payments for doctor’s
consultation fees, medication, laboratory tests and
well as macroeconomic policies. The two-way fixed hospital bills. It can be in the form of user charges
effects model shows that there have been some policy in general or cost-sharing under insurance policies
changes in the health sector among the countries [34].
over the time period. These policy changes might hap-
pen by the respective government agency in terms of
Declaration of conflicting interests
advancement of medical technology, changes in the
source of government finance and prioritization of The author (s) declared no potential conflicts of inter-
health expenditure. est with respect to the research, authorship, and/or
The result suggests that the faster movement publication of this article.
towards health financing transition would possible
through the generation of fiscal capacity by improving
ORCID
economic growth. This study has not taken into
account other measures of fiscal capacity such as rev- Deepak Kumar Behera http://orcid.org/0000-0001-6539-
enue and grants for our estimation due to the lack of 4280
Umakant Dash http://orcid.org/0000-0001-5348-9530
time series data in SEAR. Both revenue and grants
could have important sources of health sector financing
in order move towards UHC [6]. Further, generating References
additional resources through various taxation
[1] World Health Organization. Health financing strategy
measures (such as direct and indirect tax) and reprior- for the Asia pacific region (2010–2015). Geneva,
itization of health budget are important sources of Switzerland: WHO Regional Office for South-East
health sector financing, thereby it mitigates the health Asia; 2009.
10 D. K. BEHERA AND U. DASH

[2] World Health Organization. World Health Report, [21] Crivelli L, Filippini M, Mosca I. Federalism and regional
2010: health systems financing the path to universal health care expenditures: an empirical analysis for the
coverage; Geneva, Switzerland, 2010. Swiss cantons. Health Econ. 2006;15(5):535–541.
[3] Fan VY, Savedoff WD. The health financing transition: [22] Cantarero D, Lago-Peñas S. The determinants of
a conceptual framework and empirical evidence. Soc health care expenditure: a reexamination. Appl Econ
Sci Med. 2014;105:112–121. Lett. 2010;17(7):723–726.
[4] Gupta I, Mondal S. Fiscal space for health spending in [23] Khan JA, Mahumud RA. Is healthcare a ‘Necessity’or
Southeast Asia. J Health Care Finance. 2013;39(4):68–82. ‘Luxury’? an empirical evidence from public and pri-
[5] World Health Organization. World health statistics vate sector analyses of South-East Asian countries?
report. Geneva, Switzerland: World Health Health Econ Rev. 2015;5(1):3.
Organization; 2015. [24] Barros PP. The black box of health care expenditure
[6] Heller PS. The prospects of creating ‘fiscal space’for the growth determinants. Health Econ. 1998;7(6):533–544.
health sector. Health Policy Plan. 2006;21(2):75–79. [25] Sen A. Is health care a luxury? New evidence from OECD
[7] Durairaj V, Evans DB. Fiscal space for health in data. Int J Health Care Finance Econ. 2005;5(2):147–164.
resource-poor countries. World health report; 2010. [26] Cutler DM, McClellan M. Is technological change in
[8] McIntyre D, Kutzin J. World Health Organization. medicine worth it? Health Aff. 2001;20(5):11–29.
Health financing country diagnostic: a foundation for [27] de Meijer C, Wouterse B, Polder J, et al. The effect of
national strategy development; 2016. population aging on health expenditure growth: a criti-
[9] Grossman M. On the concept of health capital and the cal review. Eur J Ageing. 2013;10(4):353–361.
demand for health. J Polit Econ. 1972;80(2):223–255. [28] Sorenson C, Drummond M, Khan BB. Medical tech-
[10] Gerdtham UG, Jönsson B. International comparisons nology as a key driver of rising health expenditure: dis-
of health expenditure: theory, data and econometric entangling the relationship. Clinico Economics and
analysis. In Handbook of Health Economics 2000 Outcomes Research: CEOR. 2013;5:223.
(Vol. 1, pp. 11–53). [29] Mohapatra S, Murarka S. Improving patient care in
[11] Di Matteo L. The macro determinants of health expen- hospital in India by monitoring influential parameters.
diture in the United States and Canada: assessing the Int J Healthc Manag. 2016;9(2):83–101.
impact of income, age distribution and time. Health [30] Nisticò F, Troiano G, Nante N, et al. Socioeconomic
Policy. 2005;71(1):23–42. factors and mortality: evidences from an Italian
[12] Baltagi BH, Moscone F. Health care expenditure and study. Int J Healthc Manag. 2018: 1–6.
income in the OECD reconsidered: evidence from [31] Blanco-Moreno Á, Urbanos-Garrido RM, Thuissard-
panel data. Econ Model. 2010;27(4):804–811. Vasallo IJ. Public healthcare expenditure in Spain:
[13] Lai G. An initial investigation and analysis of health- measuring the impact of driving factors. Health
care expenditures in Hong Kong. Int J Healthc Policy. 2013;111(1):34–42.
Manag. 2017: 1–8. [32] Dickov V, Dickov A. Healthy in transition–health sys-
[14] Farag M, NandaKumar AK, Wallack S, et al. The tem in Serbia–management approach. Int J Healthc
income elasticity of health care spending in developing Manag. 2014;7(4):265–272.
and developed countries. Int J Health Care Finance [33] Reddock JR. Seven parameters for evaluating universal
Econ. 2012;12(2):145–162. health coverage: including supply-and-demand per-
[15] Schieber G, Maeda A. Health care financing and deliv- spectives. Int J Healthc Manag. 2017;10(3):207–218.
ery in developing countries. Health Aff. 1999;18 [34] World Health Organization. Global health expenditure
(3):193–205. data base. World Health Organization; 2016.
[16] Musgrove P, Zeramdini R, Carrin G. Basic patterns in [35] World Bank. World development indicators.
national health expenditure. Bull. World Health Washington, DC, United States; 2016.
Organ.. 2002;80:134–146. [36] Baum CF, Christopher F. An introduction to modern
[17] Hitiris T. Health care expenditure and integration in econometrics using Stata. Stata Press; 2006.
the countries of the European union. Appl Econ. [37] Behera DK, Dash U. Examining the state level hetero-
1997;29(1):1–6. geneity of public health expenditure in India: an
[18] Hartwig J, Sturm JE. Robust determinants of health empirical evidence from panel data. Int J Healthc
care expenditure growth. Appl Econ. 2014;46 Technol Manag. 2018;17(1):75–95.
(36):4455–4474. [38] Lu C, Schneider MT, Gubbins P, et al. Public financing
[19] Ke X, Saksena P, Holly A. The determinants of health of health in developing countries: a cross-national sys-
expenditure: a country-level panel data analysis. tematic analysis. The Lancet. 2010;375(9723):1375–
Geneva: World Health Organization; 2011. 1387.
[20] Behera DK, Dash U. Effects of economic growth [39] Palangkaraya A, Yong J. Population ageing and its
towards government health financing of Indian states: implications on aggregate health care demand: empiri-
an assessment from a fiscal space perspective. J Asian cal evidence from 22 OECD countries. Int J Health
Pub Policy. 2017: 1–22. Care Finance Econ. 2009;9(4):391.